Vous êtes sur la page 1sur 17

IRIS-Beginner Tutorial

This tutorial provides a step-by-step procedure for processing digital images obtained in raw format to yield instrumental magnitudes which are used in the next step in our tutorial: processing the data using Excel. This process assumes that, in addition to a series of sequential star field images, dark frames, flat frames have been taken. This tutorial also assumes that you have already copied the images from your camera to some directory on your computer this greatly speeds up processing!. "#"$ is a powerful image processing utility with a command-line editor and a graphical interface %&"!. 'e attempt to highlight both methods below. "n this tutorial you will 1. Initialize IRIS 2. Check raw images 3. Load and con ert images !. "er#orm Bias and $ark su%traction then &lat di ision '. (lign and stack images ). Select the green channel o# the images stack #ile *. "er#orm (+erture +hotometr,. -. .+tion/ (nal,ze each image instead stacking them (fter these steps, the instrumental magnitudes are entered into a spreadsheet which will yield calibrated magnitudes which can then be submitted to (()$*. The spreadsheet performs both: - the color correction to transform the green +$,# response to the -ohnson.s ) standard - the compensation of the differential atmospheric extinction between the stars of the image This tutorial was written by %/ara, bikeman, and #oger 0ieri with editorial changes by bkloppenborg and screenshots by bikeman. 0ote11 2ersion/ 'e recommend using the latest version of "#"$ 1.12 as of the time this document was reviewed: (ug-3455! because it implements several new features that make it easier to work with raw images. 0ote21 3la+sed Time/ 'hen processing a series of images, "#"$ will pause for some time to read, transform, calculate and save the images. The time needed for image registration could be long: several 54 seconds to minutes per image depending the image si6e and the computer. *nly the final result of these calculations will be displayed as the current image or values in *utput 7ox depending the operation. 0ote31 Linu4/ 8or ,inux users: This tutorial was tested under &buntu 54.49 using 'ine 5.:.3; with "#"$ 1.12. The only thing that does not work is the drag-and-drop of images files from 'indows Explorer. <ou have to use the =*>)E#T#(' command in order to load a series of raw images files. 0ote!1 0eed #or $ark/ +ifferential dark currents that generate a 8ixed 0attern >oise 80>! are depending of the technology level of the sensors, they are not determined by the physics fundamentals but #andom >oises: read noise, shot noise, are ?! *n recent sensors, like +$,# =/*$ 59 bits, the dark noise is negligible in exposures shorter than :4 seconds. "n that case the dark process would not apply but "#"$ doesn.t provide such option. *ne solution is to use @dummy@ images to see 9.5.3! for both bias and dark a simple constant equal to the system offset of the (nalog to +igital =onverter (+=! of the camera for the bias and @6ero@ for the dark!

0ote'/ IRIS Commands and 56I "#"$ has been developed by =hristian 7uil during many years, at the beginning it was mostly a command-line type of software, a kind of toolbox. (long years many tools have been added as new commands and some connected to a %&" but not all. Existing commandsAtools are apparently never modified but new ones are Bust added. That means older applications are not impacted by new versions: when something doesn.t work this is normally not a version issue. "#"$ is down-loadable at: "#"$ exe The definition of the newest commands is there. )arious applications examples are at end of the page. The definition of the commands up to )1.1C can be find CA3455! at: "#"$ =ommands (s tools are not all available through the %&" %raphical &ser "nterface! we will have to use both ways: the old command-line prompt and the %&". The %&" functions are accessed through the /enu 7ar, the Tools 7ar below the /enu, and some functions are accessible via a =ontextual /enu at right mouse button click. The /enu 7ar is explicit using texts, the Tools 7ar has buttons not so explicit ?

"n the following tutorial the call to the tools is shown in bold indented expressions like: 7enu Bar/ &ile8Settings... Tool Bar/ %utton with the camera +icture *r, for the commands: 9 C.023RTR(: input_file_name output_file_name number The command shall be typed after the prompt @ D @ in the =ommand 7ox when "#"$ displays that prompt if not it means "#"$ is computing and not available yet! The command name is shown in bold capital letters but "#"$ is not case sensitive! then the small italici6ed letter words are input andAor output file names you should provide at your choice, none are imposed! and values of parameters Eere number is the number of images to process!. 0ote)1 Raw $SLR Image and IRIS/ (s a +$,# doesn.t provide info on sensor temperature or controls it the dark images shall be taken during the imaging session with the same settings than the sky images. /ultiple dark images are needed to reduce their read noise by averaging and then avoid to pollute the sky images. *ne solution is to get 1 dark images before the imaging session and 1 at end, all at the operating temperature. 7ut depending the length of exposure and the sensor technology level, the dark process is not always needed To see note9 and $tep 9.3 !

"#"$ needs a master-%ias, the code being not flexible on that point To see 9.5! "f a formal bias is used it should be made from short dark exposures to avoid dark currents to develop. ( masterFbias could be reused for many sessions. &lats are mandatory to compensate the vignetting of lenses To see $tep :.:, 9.: and 9.1! ,ike for a dark multiple flat images are needed to reduce the read and shot noise by averaging. 7ut "#"$ doesn.t use flat-dark but apply only a bias to it. That means the flat images are expected taken at high enough shutter speed at which dark current can.t develop. +$,# and photo lenses having large 8*), their flats are difficult to obtain. "t.s very difficult to ensure the illumination A reflectance of the target is uniform below a couple of percents 5G is 4.45 mag ??? ! ( master-flat could be reused for many sessions. Sk, Images: ( long enough sum of exposures is needed for averaging the scintillation of the stars. ( typical series sum is about ;4 seconds but it should be noted the scintillation level is very variable depending the sky condition. The variability due to the scintillation is well quantified by calculating the $tandard +eviation $+! of the star magnitudes of a series of images or from a series of stacks. ( series of images could also be affected by @8lying $hadows@, any faint small cloud passing by, bubble of wet air... all not visible in a dark sky. To detect it a classical technique is to make several series and check the coherence of the averages of the series and relevant $+. "n case of operation without tracking eg. on tripod! each exposure would be limited to 1H51 seconds depending the focal length being used, the "$*, the stop, the defocus and the star magnitude. ( good solution is to get 1 series of 1 images then stack each series into one image using "#"$. >ext make the photometry of each of the five resulting images, then compare the results. "f one image stack strongly deviates from the others you could check the images of that series into details To see $tep 3! Then eventually eliminate that series or images! if something wrong is identified. (t end, after the spreadsheet processing, calculate the average and $+ of the 1 series, report that average, the $+ or the $tandard Error! to =iti6en $ky. Registration;Stacking: The registration of images with "#"$ could take a lot of time for images having many pixels like such of recent +$,#. This is also very depending of the registration mode being used. "f the lens is a 6oom there is a risk of serious geometry distortion of the image. "n this case the @%lobal matching@ and @Iuadratic@ options shall be used but the registration takes a lot of time. "f a single long focal lens is used the distortion is probably very low. Then each series should be shot quickly, doing it within one minute should keep the star field rotation low enough. "n this case the @*ne star@ mode should be usable and much faster. "f there is a significant mismatch of the stars at the periphery either @Three 6ones matching@ or @%lobal matching@ with @(ffine@ should be applied. @=ubic@ is to be used in case of strong distortion. (nother way than stacking, when few stars are involved, is to make the photometry of each image and then average the results. 8urther it.s possible to make a dee+er statistic anal,sis. This process is described under the $tep C option. IRIS #ile series num%ering/ "n all cases the numbering of file sequences is generated A imposed by "#"$. The way this numbering is done is not compatible with most camera format and anyhow it starts at 1 #or each new se<uence, input or output. The format is fileFname5, file-name54, fileFname544... without leading 6ero. "f file series already exist in the working folder having the name you use for a new operation they will be erased by the new ones. This is the reason why a camera file series shall be loaded with the copy-past process of the @+ecode raw files...@ box. "n other cases eg. ,inux! the only solution is to change the numbering of the input files. This can be done with software you would find on the web.

Links/ =hristian 7uil.s "#"$ 8rom where to download "#"$ and find a lot of information "#"$ =ommands The documentation of "#"$ commands up to version 1.1C pdf! +$,# 'ikipedia, general presentation with many links and table of sensor si6es $ensor Types /ore on sensors, $i6e, $># vs si6e... =8( A 7ayer =olor filters used in +$,# "mage >oise 80>, read, shot, quanti6ation... with links to more details %amma The standard non-linear coding of image luminance -0E%, s#%7, >T$=...! +ave =offin.s $ite +=#(', the raw image file converter used by many imaging software dcraw.exe =hoice of dcraw.exe compilations for various systems =$ky $preadsheet To finish the processing, color transform and extinction compensation =alibration $tandard ,ist of =omparison stars and related data

Step 1 - Initialization: 1.1 Files Settings: 7enu Bar/ &ile 8 Settings... =hoose: - your working path where you have =*0"E$ of your raw images files!, - the script path if you want to use some scripts.. ! and - the file type for images produced by "#"$ eg. 0"= !.

1.2 Camera Settings: Tool Bar/ %utton with the camera +icture *pen =amera $ettings and leave the default values except: - 7inning 5x5, - =amera model select appropriate! - #aw interpolation method @,inear@, - 'hite balance >*T selected. >ote: @=amera model@ is used to determine the color order of the =olor 8ilter (rray and the suffix extension! of the raw image files then other types are not recogni6ed by "#"$! /ost =anon are #%%7, >ikon 7%%#. This information is available as output of the +=#(' raw image file converter. "f your camera is not in the list choose first a similar one from the same brand. "f the problem remains visit the site of +ave =offin.s dcraw.exe the converter used by "#"$! &sing @dcraw -i -v "/%Fxxxx.=#3@ =anon! as a 'indows command shows the image info including the =8( order. >ear 144 camera types are covered.

Step 2 - Checking Images: "#"$ has some simple graphical tools that enable us to check images quickly before processing them. This is useful for optimi6ing the shooting parameters, framing, avoiding saturation... "t could be used for further analy6ing the images when problems occur. The raw images can be visuali6ed simply as a =8( A 7ayer image using the raw loading: 7enu Bar/ &ile 8 Load a R(: #ile... Slider Bo4/ (uto %utton $lider 7ox: #ange option! #aw images are coded linearly on 53 or 59 bits and are not in a shape to be properly visuali6ed. &sual images like Bpg! are coded, shaped, with a low power function not far a log.. ! following the standard 7T J42 to be properly displayed %amma!. 'ithout such standard gamma raw images look very dark on usual displays, only the brightest stars are visible if not saturated. Even applying a gamma would not be enough as our images go very deep, fainter stars recorded are ; magnitudes below the brightest non-saturated, this is a ratio of 314, too much for any display. "#"$ provides a tool that enables us to properly clip the image at an adBustable level and also control the black level. This is the @Threshold@ $lider 7ox that pop up at start of "#"$ if not use: Tool Bar/ Button with two arrows! - The upper slider sets the clipping level of the image, - The bottom slider adBusts the black level. The @(uto@ button starts the automatic search of the background level of the image and applies clipping and black level shift defined through @#ange@ The @Threshold #ange@ values are such of the +$,#: - 9421 A 4 for a 53 bits - 5;:C: A 4 for a 59 bits. @+elta Ei@ determines the clipping level above the background plus offset. @+elta ,ow@ elevates the display black level by its value, this better shows the faint stars and the noise. 8or our 53 bits samples +elta Ei K 144 and +elta ,ow K 544 is fine. Then during your image checking you could change the clipping and black levels using the sliders at will. (ll this onl, a##ects the dis+la,1 not the image data.

There are other #unctions under/ 7enu Bar82iew8... that could %e used to either generate a standard gamma1 a logarithm or an, other ad=ustments %ut all are a++lied to the data and shall ne er %e used #or +hotometr,.

2.1 Checking Saturation,

ange, !ackgroun", #oise...

(nother simple tool is available through the mouse. "f you want to check a star Bust draw a box around it with the left button down, then click right, a dialog box pops up, click @$tatistics@ (nother box shows you various parameters of interest. 7ouse >9 draw %o4 >9 click right >9 dialog %o4 >9 Statistics Saturation check/ The @/ax@ value is such of the highest pixel. 'ith a ==+ 53 bits with antiblooming we should probably not go higher than 3144H:444. "n a 59 bits =/*$ the sensor would saturate about 59444 (+&s at "$* 544 and then, at higher "$*, the (+= range should clip the image at 5;:C: (+&s but much less electrons, limiting the dynamics... ! Background Le el/ The @/edian@ value shows it. To get an exact value the box should be drawn in a dark area. Then the @/ean@ is a more accurate value. 'ith =anon +$,# that level includes the system offset of the (+=, in general 53C for 53 bits and 5439 for 59 bits. That value shall be subtracted from the image data to determine the true background level. 0oise le el/ 'hen used on a dark the @sigma@ shows the noise level %aussian read noise and possibly some dark current impulse noise! "f used on images it could includes some other signals. 2.2 Checking $efocus, %rail 7ouse >9 draw %o4 >9 click right >9 dialog %o4 >9 ? ..Cro+/ a box drawn around a star and then using the 6oom button two buttons before @L5@ ! permits to visuali6e the star pixels into details, to check the defocus, the trail. This is a @7ayer@ image, the #%%7 sensor arrangement, but in 7M'. "f you point a pixel with the mouse its value is shown at the bottomright of the window after its coordinates. <ou can analy6e into details a saturation, pixel involved, color involved...

..Sha+e/ is another way to visuali6e a saturation, a defocus...

2.& Setting the 'hotometr( )perture 7ouse >9 draw %o4 >9 click right >9 dialog %o4 >9 5rowth cur e This function shows the photometry error in function of the si6e of the inner circle of the photometry tool. "n our test set the photon count is completed below 4.45 mag for a 53 pixels radius. ( ; pixel radius would result in losing photons corresponding to an error of 4.3 mag or 5JG of the flux. To be applied in $tep J.

2.* Checking the "ark images This can be done with the @Eistogram@ function applied to a raw dark image: 7enu Bar/ Load a Raw #ile... 7enu Bar/ 2iew 8 @istogram "lot 7enu Bar/ .+tion 8 (4is Setu+

Step & - +oa"ing an" Con,erting a- Image Files: &.1 .pening the loa"ing !o/ 7enu Bar/ $igital +hoto 8 $ecode R(: #iles... This will open up a box in which files should be dragged-and-dropped from 'indows Explorer. $rag and dro+ ,our raw image #iles eg. for =anon: "/%F>>>>.=#3 files! here. &.2 Con,ert ra- image files to CF) files Bo4 Button/ AB C&(A This dialogue lets you specify an output generic name. (fter dropping in your images, press the @--D=8(@ button to decode the images ,inux users see the note here below!

&.& )ppl( this process for: - star field images e.g. use the name @img@!, - dark frames suggested name @dark@!, - flat field images suggested name @flat@! and - bias frames @bias@ optional!. >ow you should have a series of images, darks, flats and optionally bias frames in the working directory at the format you configured in 5.5 >ote: "f you are running "#"$ under ,inux also works under 'indows!, open up the command line input dialog box, the button to the left of the camera button, and type at the prompt @D@ 9 C.023RTR(: input_file_name output_file_name number 'here @input_file_name@ is the image file generic name from your camera and @output_file_name@ is the generic names of files that "#"$ creates with the =8( order defined by 5.3 and file type defined in 5.5 7oth names do not include the sequence numbering imposed by "#"$! nor the file suffix. The full ,inuxA'indows name of input files would be: @"/%F422>.=#3@ for =anon for example, where @>@ is a series of numbers that starts at @5@ and ends at @number@ without leading 6eros. This is in general not compatible with the camera numbering format and you would have to convert the numbering of your files.

Step * - 'reprocessing 0 !ias, $ark, Flat: &nlike ==+ cameras, +$,# cameras don.t provide the user with the ability to control the sensor temperature and apply a previously computed master dark and bias combination to the images. 7ecause of this, one should take dark shots during the imaging session which will include the bias pattern of the camera. (lthough the +$,# dark frames will account for the bias, "#"$ requires separate bias offset! frames. 7ecause of this, we can either create a formal bias frame or a fake @dummy@! bias frame, both work equally well, but one of the two must be completed. (s said in >ote9 recent sensors have very low differential dark currents and bias pattern, for exposures shorter than :4 sec dark and in any case bias! can be replaced by dummies. *.1 !ias1.ffset Frame *.1.1 Creating a Formal !ias To generate a 7ias /aster 8rame go to: 7enu Bar/ $igital "hoto 8 7ake an o##set... - Type the appropriate generic name the one introduced in $tep :, e.g. .bias.! - The number of bias images, Bo4 Button/ .C - Then select the @command line@ button the one on the left from @camera@ button! and type at the prompt: 9 S(23 master-bias or any name you would like! or 7enu Bar/ &ile 8 Sa e.. The master-bias kee+s alid for a given +$,# as long as its electronics remains in good condition, a check every couple of months would be enough. "t is "$* depending.

*.1.2 Creating a 2"umm(2 !ias $ome +$,#A+$= cameras use a @system offset.@ This offset is built into the camera.s analog to digital convert (+=!. This offset will take the black level ideally 4! and encode it as a higher value usually 53C for 53-bit cameras, 31; or 5439 for 59-bit cameras!. the result is that generated values can be less than the offset 6ero position or, in other words, negative brightnesses are permitted? eg. %aussian read noise! "#"$ doesn.t handle this parameter, even if the camera types are identified, therefore we need to determine it for generating the artificial bias. The easiest way to do this is to load one of the dark frames and look at its black level. To measure the black level use the command prompt box and type: 9 ST(T The output window will show an average value of the whole image. &se this value as the black 6ero-point reference. "f you would rather not use the command prompt, you can select a blackened portion of the image using the mouse. To do this left click on the area, draw a box, right click and select @statistics.@ "#"$ will show the average value in the boxed area. &se this value as the black 6ero-point.

>ow we create the artificial bias frame. &se the command prompt box and type: 9 &ILL ,alue 'here @value@ is the black 6ero-point from above. >ow, save the master bias by typing: 9 S(23 master-bias $uch dummy master-bias is valid for ever. <ou are done. !.2 3aster $ark: >ow we need to create a master dark: 7enu Bar/ $igital "hoto 8 7ake a dark... Type in the box the generic name, the offset image the master bias frame you Bust created!, and the number of images in the dark sequence. $elect either the mean or median method they work equally well, median has the property to eliminate any influence from extreme values!, and press *N. Then type the command: 9 S(23 master-"ark or choose: 7enu Bar/ &ile 8 Sa e... To create a Adumm, darkA if your +$,# is recent and your exposure short enough, Bust use the same process as in 9.5.3 but apply 9 &ILL D *.& 3aster Flat: The flat process is mandatory for eliminating the vignetting of the lens. )ignetting as high as :4H94G are common at large aperture and could induce large differential magnitude errors depending location of the reference star and the target star. #epeat the same process as used for darks: 7enu Bar/ $igital "hoto 8 7ake a #lat-#ield... Except before saving type at the prompt: 9 5R3EF&L(T in the command box to normali6e the red and blue =8(-flat. Then save the file as @master-flat.@

That master-#lat kee+s alid each time you use the same +$,# and ,ens at the same aperture, same focus, same focal length for a 6oom. =heck that no large dust has polluted the low-pass filter that is few millimeters in front of the sensor, this is the only level a dust could be an issue if you use a large lens aperture 8A9 or less! (ny flat would not correct it properly. *.* 4ot 'i/els $etection an" ecor"ing: This function automatically detects hot pixels above a threshold level. $uch defective pixels couldn.t be processed properly through dark or flat. Their coordinates are recorded in a file for further processing of the images replacement by surrounding average! "t might take some experimentation to determine the threshold value you need, but a good starting value at "$* 544 is about 344 (+&s for a 53-bit ==+ cameras and 144 for a 59-bit =/*$ cameras offset removed!. That threshold should be more or less proportional to the "$* being used. The number of hot pixels could be 14 in the first case or 54 in the second very depending of the sensor quality class. ,oad the processed master dark frame created in $tep 9.3 , offset removed by "#"$! and type the following at the command prompt: 9 &I0$F@.T cosme number 'here @number@ is the threshold value you have selected and @cosme@ is the name of the file where the results are recorded you could chose any you like!. =heck the *utput 7ox it should open automatically!. "f there is a number of hot pixels not detected, change the threshold accordingly. "t.s possible to differentiate the random noise, the impulse noise and the hot pixels using the histogram function, the threshold should be well above the %aussian >oise distribution, then the stack of +ark "mpulses and below the isolated peaks on the right corresponding to the hot pixels. <ou can estimate their number from the graph. 7enu Bar/ 2iew 8 @istogram option: adBust the axis to 9444 (+&s or so, density 54 ! !.5 Finish 'reprocessing %o to: 7enu Bar/ $igital "hoto 8 "re+rocessing.... - "nput the generic name of you data images img! - the offset map name master-bias!, - dark map name master-dark!, - flat-field map name master-flat!, and - the cosmetic file name cosme!. - an output generic name e.g. .img-cal.! and - the number of image files.

+* >*T tick dark optimi6e as it will take longer to finish!.

Step 5 - )lignment of Images: 5.1 CF) con,ersion to 6!: 7enu Bar/ $igital +hoto 8 Se<uence C&( con ersion... and select - the generic name that you gave at $tep 9.1 .img-cal.! - give an output name .img-cal-conv.! and - number of your star images and - select =olor *utput files type . 6o to Step 7 if (ou prefer to make the photometr( of all (our images an" not stack them 8#ote9: 5.2 6enerate egistration Coor"inates %o to: 7enu Bar/ "rocessing 8 Stellar registration and put: - the sequence name of 1.5 .img-cal-conv.!, - give an output generic name e.g. .img-reg.! and - the number of images - choose @%lobal matching@ and @Iuadratic@ transformation "f the @*ne star@ option is used, >ote;, you shall load first the first image and draw a box around the star!

5.& Stacking the images: 7enu Bar/ "rocessing 8 (dd a se<uence... - give the input generic name of 1.3 img-reg! - the number of images, and - choose arithmetic addition, and - normali6e if overflow D :3J44 ! Sa e the image G eg. imgFstk !

Step 9 - Selecting the green channel: The recent versions of "#"$ i.e. 1.1J or newer! include a quick and easy way to pull out the #%7 channels: Load the stacked image #ile eg. imgFstk ! if not already in memory and go to: 7enu Bar/ $igital "hoto 8 R5B Se+aration... Then Bust type in the dialog box: - the names of the color channel files: e.g. final-r, final-g, final-b! - =lick *N <ou.re done separating the #%7 channels? Step ; - 'hotometr(: The coordinate system of "#"$ places its 6ero-point 4,4! at the bottom left corner of the screen instead of the top left used by most imaging applications but the images are normally oriented. Load the green image #ile #rom ste+ ) e.g. finalFg ! and look around for a while. "n our sample data =apella is prominent and saturated. The isosceles triangle formed by the %ot (sterism is a little below at its right, near vertical, easy to spot. Eta and Oeta (&# form the short base of the triangle and Epsilon is at the top, the sharp angle. <ou can adBust the visibility of the stars by clicking @(uto@ with the @#ange@ settings shown at below. /oving the upper slider in the Threshold 'indow to see $tep 3! is a further possibility.

These dis+la, settings do not a##ect the data of the image used for photometry To see $tep 3, other @/enu7arP)iew..@ functions change the data ? !

;.1 'hotometr( Setup 7enu Bar/ (nal,sis 8 (+erture +hotometr,... 'ith our sample data, you should probably use the si6e of the reading apertures shown at left, but for your own data you may need to change these si6es so you may wish to review how to si6e a photometry aperture To see 3.:! The inner circle is the area in which the star photoncount is done, the external annulus is used to measure the average or median level of the background that is next subtracted from the inner circle pixel values. The intermediate annulus is used as a guard to avoid the halo of the star pollutes the background. "t could be used to eliminate a nearby companion star. The inner circle should well intercept most signal from the star and by the way be significantly larger than the largest visible star footprint. The same settings shall be used for all stars of a given image. The background area should be large to minimi6e the involved noise if no background star fall in it. The @/edian 7ackground@ method is less sensitive to faint background stars pollution but is less accurate integer value! than the mean floating point value! ( small rounding error due to the integer mode will be multiplied by the number of pixel in the inner circle. *ne half (+& error would result in 331 counts deviation of the star intensity in our example ? <ou should choose median - or not - depending the background conditions. "f the catalogs do not show a risk of faint star pollution in the area don.t use median background. "f you carefully center the circles to any star and click left you will take an output in the *utput 7ox with values for "ntensity and /agnitude along with other parameters, coordinates...!. The important one is the /agnitude or "ntensity which are equivalent!. "f you would like a tentative estimation of the magnitudes you have to set the /agnitude =onstant The instrumental Oero-0oint! This is optional, you can also transfer the instrumental results without 6eropoint correction to the spreadsheet for further color and extinction corrections. <ou would set that /agnitude =onstant on the primary =omparison $tar. "n our case it.s probably Eta (&# which has a recommended (()$* magnitude of :.5J3 ;.2 'erform the )nal(sis: 7enu Bar/ (nal,sis 8 (+erture "hotometr,... - =ircle number K : - $et the circles radius - /agnitude =onstant K 4 =enter the circles on the primary comparison star, click left, get the instrumental magnitude from the *utput 7ox, calculate the /agnitude =onstant. 7agnitude Constant > Catalog 7agnitude - Instrumental 7agnitude signed!

%o to: 7enu Bar/ (nal,sis 8 7agnitude Constant... Set the Constant1 .C >ow all readings will be made relative to that =omparison $tar but neither color nor extinction corrected, there are only approximate and provisional. <ou could check that Oeta (&# gets to some :.C2J instead :.J11 (()$*! This is due to the color difference between the very blue Eta (&# and the well red Oeta. Q/ag K N x Q 7-)!... here N H 4.545 and Q 7-)! is the =olor "ndex difference between Oeta 5.33! and Eta -4.5C!

>ow you should apply the process repetitively to all involved stars of each stacked image or each image if you do not stack! These stars are your variable target plus the @ensemble@ of comparison stars needed for the process implemented in the spreadsheet =olor transform and extinction gradient compensation! 8our to six comparison stars should be enough, when possible more is better... $omeones are also using @check@ stars: Bust apply the same process. The results are recorded in sequence by "#"$ into the *utput 7ox for all images. That list of results can be saved as a @xxx.dat@ file a simple ($="" text file! using: .ut+ut Bo4/ &ile 8 Sa e as... There is no possibility to add comment or name, it.s better to use the same sequence for each stacked image to not mix the results. Then you could copy and past each instrumental magnitude into the spreadsheet for further processing. &nder best condition and enough images a final accuracy of a couple of mmag being possible, it is recommended to transfer magnitudes with three decimal places. EouHre doneG >ow that you have obtained instrumental magnitudes, you need to compute calibrated magnitudes. The +$,# +ocumentation and #eduction team has setup an Excel $preadsheet to help you finish your analysis.

Step 7: 'hotometr( of all images -ithout stacking 8.ption: "f you prefer to make the photometry of all your images and not stack them, Bust after the point 1.5 apply the following sequence #%7 separation: 7enu Bar/ $igital "hoto 8 Se<uence R5B Se+aration... Then Bust type in the dialog box: - the generic name of the #%7 images from 1.5! - the names of the color channel files: e.g. final-r, final-g, final-b! - the number of images to process. - =lick *N <ou.re done separating the #%7 channels? >ow continue the process at the $tep J and repeat the photometry operation for each image. Then you could calculate the average and $+ of the instrumental magnitudes of all images or group them to form several series! and next report it to the spreadsheet. RRRRRRR